Data-driven approaches enabling the design of community energy systems in the Global South

This thesis answers the primary research question: What techniques can be developed to address the challenges of context specificity and replication at scale in the design of electricity access systems suitable for rural, low-income communities in the Global South? This is addressed through two sub-questions: 1. To build electricity access systems that are suitable for specific contexts, how can we take account of end-user needs, aspirations, and context in the design process? 2. To support replication of electricity access systems at scale, how can we use time series electricity demand data coupled with demographic data to estimate electricity demand for these communities? Five techniques are developed. The first is the Service Value Method, which gathers and integrates end-user needs, aspirations, and context into engineering design practice. It incorporates data-gathering methods suitable for Global South contexts. The remaining four techniques relate to how electricity and demographic data can be used to estimate electricity demand for new communities, supporting replication of systems at scale. Time series electricity data are linked to household demographic data, and exploited to estimate a) daily electricity consumption (relevant for battery-charging systems), and b) daily load profiles (relevant for micro-grids). For each demand metric, the approach is to retain the variation in electricity use for each household across different days. Throughout this thesis a case study from Kenya of lighting electricity data from households using two solar nano-grids is used to demonstrate these techniques. Results from the case study, comprising a dataset of household lighting electricity data, and the typical use patterns derived from them, contribute to demand estimation research for these contexts.